Back to results
US 12,597,488 B2Granted
Use of machine learning models for prediction of clinical outcomes
AssigneesThe Henry M. Jackson for the Advancement of Military Medicine, Inc., The United States of America as represented by the Secretary of the Navy, Duke University Medical Center, Emory Univeristy
InventorsSeth A Schobel, Vivek Khatri, Felipe Lisboa, Matthew J. Bradley, Christopher J. Dente, Timothy Buchman, Allan D. Kirk, Jonathan A. Forsberg, Eric A. Elster
Life of this patent
Estimating the term…
Abstract
The present disclosure describes methods and systems for predicting if a subject has an increased risk of having or developing one or more clinical outcomes, including prior to the detection of symptoms thereof and/or prior to onset of any detectable symptoms thereof. The present disclosure also describes a method of generating a model for predicting one or more clinical outcomes.
Claims22
A method of generating a model predicting one or more clinical outcomes for a subject, comprising:
- generating, by a processor, a training database storing first values of a plurality of clinical parameters and clinical outcomes associated with a plurality of first subjects;
- executing, by the processor, a plurality of variable selection machine learning models to select sets of model parameters from the plurality of clinical parameters for each variable selection machine learning model, wherein each set of the model parameters from the plurality of the model parameters comprises a different subset of the clinical parameters;
- associating, by the processor, a plurality of classification machine learning models with the sets of model parameters, wherein each classification machine learning model of the plurality of classification machine learning model is associated with a set of model parameters of the sets of model parameters;
- executing, by the processor, each one of the plurality of classification machine learning models and the associated set of model parameters to generate predictions of one or more clinical outcomes;
- calculating, by the processor, a performance metric comparing a clinical outcome prediction of each combination of one of the plurality of classification machine learning models and a selected set of model parameters to one or more clinical outcomes, wherein the performance metric comprises a difference in Kappa score, a difference in sensitivity, and a difference in specificity;
- selecting, by the processor, a respective combination of a candidate classification machine learning model and set of model parameters having a highest Kappa score, a highest sensitivity, and a specificity over a threshold;
- and
- outputting, by the processor, a plurality of models for predicting one or more clinical outcomes, each model comprising a selected combination of candidate classification machine learning model and set of model parameters.
5 more elements
A method for predicting one or more clinical outcomes for a subject comprising:
- receiving, by a processor and from a second subject, a second value of at least one clinical parameter of a plurality of clinical parameters;
- executing, by the processor, a pre-trained model for predicting one or more clinical outcomes of the second subject using the second value of at least one clinical parameter, wherein the model is pre-trained by performing operations comprising:
- generating a training database storing first values of the plurality of clinical parameters and one or more clinical outcomes associated with a plurality of first subjects;
- executing a plurality of variable selection machine learning models to select sets of model parameters from the plurality of clinical parameters for each variable selection machine learning model, wherein each set of the model parameters from the sets of model parameters comprises a different subset of the clinical parameters;
- associating, by the processor, a plurality of classification machine learning models with the sets of model parameters, wherein each classification machine learning model of the plurality of classification machine learning model is associated with a set of model parameters of the sets of model parameters;
- executing each one of the plurality of classification machine learning models the associated set of model parameters to generate predictions of the one or more clinical outcomes;
- calculating a performance metric comparing a clinical outcome prediction of each combination of one of the plurality of classification machine learning models and a selected set of model parameters to one or more clinical outcomes, wherein the performance metric comprises a difference in Kappa score, a difference in sensitivity, and a difference in specificity;
- selecting a respective combination of a candidate classification machine learning model and set of model parameters having a highest Kappa score, a highest sensitivity, and a specificity over a threshold;
- outputting a plurality of models for predicting the one or more clinical outcomes, each model comprising a selected combination of candidate classification machine learning model and set of model parameters;
- and
- outputting, based on executing the second value with the pre-trained model, a predicted one or more clinical outcomes of the second subject.
8 more elements
Dependencies are inferred from the claim text.